33. Gauss Theorem Theorem 33.1 (Gauss Theorem).

Let M R3 be a smooth 3-manifold with boundary, and let F : M R3 be a smooth vector eld with compact support. Then div F dx dy dz =
M M

F dS,

where M is given the outward orientation. Example 33.2. Three point charges are located at the points P1 , P2 and P3 . There is an electric eld E : R3 \ {P1 , P2 , P3 } R3 , which satises div E = 0. Suppose there are four closed surfaces S1 , S2 , S3 and S4 . Each Si divides R3 into two pieces, which we will informally call the inside and the outside. S1 and S2 and S3 are completely contained in the inside of S4 . The inside of S1 contains the point P1 but neither P2 nor P3 , the inside of S2 contains the point P2 but neither P1 nor P3 , and the inside of S3 contains the point P3 but neither P1 nor P2 . The inside of S4 , together with S4 , minus the inside of S1 , S2 and S3 is a smooth 3-manifold with boundary. We have M = S1 S2 S3 S4 .

Recall that primes denote the reverse orientation. (33.1) implies that E dS
S4 S1

E dS
S2

E dS
S3

E dS E dS
S3

=
S4

E dS +
S1

E dS +
S2

E dS +

=
M

E dS div E dx dy dz
M

= = 0.

In other words, we have E dS =

S4 S1

E dS +
S2

E dS +
S3

E dS.

Proof of (33.1). The proof (as usual) is divided into three steps.
1

Step 1: We rst suppose that M = H3 , upper half space. Suppose that we are given a vector eld G : H3 R3 , which is zero outside some box K = [a/2, a/2] [b/2, b/2] [0, c/2]. We calculate: c b a G1 G2 G1 div G du dv dw = + + du dv dw u v w H3 0 b a
c b

=
0 c b a

(G1 (a, v, w) G1 (a, v, w)) dv dw (G2 (u, b, w) G2 (u, b, w)) du dw

0 b a a

+ +
b

(G3 (u, v, c) G3 (u, v, 0)) du dw

a b a

=
b a

G3 (u, v, 0) du dw.

On the other hand, lets parametrise the boundary H3 , by g : R2 H3 , where g (u, v ) = (u, v, 0). In this case It follows that G dS =
( H2 ) R2 b a

g g = = k. u v du dv Gk G3 (u, v, 0) du dv.
b a

= Therefore G dS =
H2 ( H2 ) b a

G dS G3 (u, v, 0) du dv.
b a

= Putting all of this together, we have

div G du dv dw =
H3 H2

G dS.

This completes step 1.

2

Step 2: We suppose that there is a compact subset K M and a parametrisation g : H3 U M W, such that (1) F (x) = 0 for any x M \ K . (2) K M W . We may write g (u, v, w) = (x(u, v, w), y (u, v, w), z (u, v, w)). Dene G : H3 R3 , by (y, z ) (x, z ) (x, y ) F2 + F3 (v, w) (v, w) (v, w) (y, z ) (x, z ) (x, y ) G2 = F1 + F2 F3 (u, w) (u, w) (u, w) (y, z ) (x, z ) (x, y ) G3 = F1 F2 + F3 , (u, v ) (u, v ) (u, v ) G1 = F1 for any (u, v, w) V and otherwise zero. Put dierently, G(u, v, w) = F A if (u, v, w) U 0 otherwise,

where A is the matrix of cofactors of the derivative Dg . One can check (that is, there is a somewhat long and involved calculation, similar, but much worse, than ones that appear in the proof of Greens Theorem or Stokes Theorem) that div G = div F det Dg = div F We have div F dx dy dz =
M H3

(x, y, z ) . (u, v, w)

div G dx dy dz G dS,
H2

= =
M

F dS,

where the last equality needs to be checked (this is relatively straightforward). This completes step 2. Step 3: We nish o in the standard way. We may nd a partition of unity
k

1=
i=1

i ,

where i is a smooth function which is zero outside a compact subset Ki such that Fi = i F is a smooth vector eld, which satises the hypothesis of step 2, for each 1 i k . We have
k

F =
i=1

Fi .

and so
k

curl F dS =
S i=1 k S

curl Fi dS Fi ds
i=1 M

= =
M

F ds.